Bulk Aluminium Panel Curtain Walls for Slab Movement and Joint Design

Aluminium panel curtain walls have to absorb 28 mm of movement on a 6 m bay, 20 mm of it from the slab and 8.3 mm from the mullion, and that is 8 mm more than the 20 mm stack joint most facade schedules still carry. A joint that size is undersized by the width of a pencil, on every floor of the building.

How much movement does an aluminium panel curtain wall actually absorb?

Three sources, and they add up; none of them cancels another. The slab deflects under its own long-term load and again under live load. The aluminium moves with temperature about seven times faster than the glass set in front of it. The frame is erected to a tolerance that has to sit somewhere inside the same joint. A facade that counts only the first of the three is the one that splits a sealant bead in the second summer.

The order of magnitude is worth carrying in your head: 6 m of bay, 20 mm of slab, 8 mm of mullion, 5 mm of glass differential. Nothing on that list is unusual. What is unusual is a 20 mm joint asked to take all of it at once.

The panel band changes one thing in that arithmetic. A 3 mm aluminium composite panel is stiffer than glass of the same size, so it refuses to follow the mullion as the mullion bows, and the seal between panel and mullion takes the difference as shear. That is why panel systems usually want the movement joint at the transom sitting on the slab line, where the panel band already breaks. The smaller version of the same problem on a punched opening is set out under architectural aluminium windows.

What does a 60 K temperature swing do to 6 m of mullion?

Delta L equals length times coefficient times temperature change. For 6063-T5 aluminium the coefficient is 23.1 x 10^-6 per kelvin, so 6,000 mm at 60 K gives 6,000 x 23.1 x 10^-6 x 60 = 8.3 mm. The 60 K is worth defending: a dark powder coated or anodised mullion on a south elevation reaches roughly 75 C in summer sun and is installed at about 15 C, and that 60 K gap is what the joint sees.

The glass in the same opening moves 3.2 mm across the same 60 K, because soda-lime glass expands at 9 x 10^-6. The 5.1 mm difference is what the glazing detail has to swallow, 0.85 mm per metre of height, and it explains two details that look arbitrary until you do the sum: glass sits on two setting blocks at the quarter points, and the bite stays at 15 mm or more on a tall unit.

On a full height structural silicone joint the same difference is why the joint is sized in shear. A silicone bite that carries the glass weight plus 0.85 mm per metre of differential movement ends up wider than the one that only carries weight, and a facade quoted on the weight only number will show a thin white line at the joint within a few years as the silicone starts to fail at the edge.

Why is the slab the larger half of the budget?

Because an aluminium panel curtain wall hangs on the slab edge, so every millimetre the slab drops arrives at the head of the mullion as a vertical translation. A 6 m span checked at L/300 gives 20 mm of long-term deflection once creep and shrinkage are included, and the same span at L/360 under live load gives 16.7 mm. The two do not add, since the long-term case already covers the service condition, but the larger one sets the joint.

Twenty millimetres across a 1.2 m module is a slope of 1 in 60. Over ten bays the transom line steps 20 mm and the eye reads it as a bend, even though every individual panel is flat and square. This is the failure buyers report as "the facade looks wavy" and it is a slab stiffer than the joint, in that order.

Slab figures are the L/300 and L/360 limits applied to a 6,000 mm span; the mullion figure uses 23.1 x 10^-6 per kelvin and 60 K; the glass figure uses 9 x 10^-6. These are calculated from the span and the material coefficients, not measured on a built facade. Calculated, not measured.

Where does the movement show up first on a finished facade?

At the head, in the sealant. A class 25 sealant works within plus or minus 25 per cent of its joint width, so a 20 mm joint gives 5 mm of movement each way. The slab alone asks for 20 mm one way. The bead does not fail because the sealant was cheap; it fails because it was asked for four times the movement it can take, and the first symptom is a cohesive tear along the bond line on the sunny elevation of an aluminium panel curtain wall.

The second symptom is quieter: gaskets walk out of their grooves at the transom, a few millimetres a season. The third is the expensive one, glass bearing on the rebate at a corner where the mullion has bowed inward. All three arrive in the same order on every facade that is short of joint, and the inspection that catches them early is a straight edge held against the transom line, not a hose test. The same sequence on a smaller scale is described under windows for large buildings.

What should the facade schedule fix before the order?

Five lines, each a number on the drawing and none of them a note. Module width and height, because the joint budget is per bay. Stack joint nominal width together with the movement it can absorb, stated as a pair. Anchor slot length with the bolt set at mid slot, so the slot can move both ways. Glass bite and setting block positions. Sealant movement class, written against the joint width it has to serve.

Send the elevation with the module sizes and the structural span to kristin@haolvbuilding.com and the movement budget comes back bay by bay, with the stack joint each one needs. That conversion is worth doing before the slab is poured, because a 6.5 m span and a 6.0 m span are 22 mm and 20 mm of deflection, and only one of them fits a 25 mm joint with any margin left.

A habit that costs nothing: put the movement table on the facade drawing as four rows, one per source, with the total underneath. Fabricators price what they can see, and a joint that appears in the schedule as a number tends to survive value engineering better than one that appears as a detail reference. The same discipline across a whole envelope, including the entrance line, is set out under commercial building glazing.

Which specification holds this movement budget?

Read the table as a build sheet for an aluminium panel curtain wall. The module follows the slab grid. The mullion depth follows the wind load and the deflection limit. The bite follows the movement, and the joint follows everything above it. Where a 50 mm sightline is wanted, the answer is a deeper section, and a thinner wall leaves the mullion bowing into the joint it was supposed to protect. At the base of the same elevation the movement lands on a door head, and that junction is set out under aluminium entrance doors.

Usual range for this system family. Wind load, module and joint are project values confirmed against the structural drawings; the U-values are area weighted from standard glass and frame figures, calculated, not measured.

Frequently asked questions

What is the minimum order for an aluminium panel curtain wall and how long is production? Minimum order is 10-20 units and lead time is 25-30 days from approved shop drawings. Shop drawings come first on a facade, because the module and the stack joint have to be agreed before any profile is cut.

How is a curtain wall quoted, per square metre or per bay? Per square metre of finished facade, with the module, the joint width and the glass build stated, so two quotes can be compared line by line. A price per bay hides the module and is where most of the difference between suppliers sits.

Can the stack joint be widened on site if the slab turns out to deflect more? Up to a point. The joint has a nominal width and a movement capacity, and the two travel together: widening it by 5 mm buys 5 mm of travel only if the sealant and the gasket are rated for the new width. Past that, the cover caps no longer reach and the answer is a shimmed sub-frame, which is a workshop job.

Does a wider joint leak more? No, a drained and ventilated system leaks less with a wider joint, because the cavity behind the joint is what equalises pressure. Water arrives at the inner line only when the cavity cannot drain or cannot breathe, which is a function of the weeps and the baffle, and of the joint width only in the sense that a joint too narrow to fit a baffle has no cavity at all. On an aluminium panel curtain wall that cavity is the component doing the work, so the joint detail and the weeps are one item on the schedule.

Ordering from Haolv

Haolv Building Materials Co., Ltd. has been making doors, windows and facade systems for 19 years, aluminium panel curtain walls among them, and returns a module and joint proposal against the elevation before a price is fixed. Certification for the profiles and the glazing is available on request with the quotation. Minimum order is 10-20 units and lead time is 25-30 days.

Send the elevation and the structural spans to kristin@haolvbuilding.com, or start from Haolv Building Materials Co., Ltd. and the movement budget, the stack joint and the panel module come back together with the quotation.

What this guide claims and what it does not: the 8.3 mm, the 5.1 mm and the 20 mm are arithmetic on stated spans and published coefficients, so anyone can redo them with a calculator. The deflections depend on the structural design of a specific building and are confirmed against its own drawings; the sealant class and the gasket are confirmed against the joint width actually built. None of the numbers here come from a measured facade.

References

1. EN 13830:2020 — Curtain walling. Product standard.
2. EN 12152:2002 — Curtain walling. Air permeability. Performance requirements and classification.
3. EN 12154:2000 — Curtain walling. Watertightness. Performance requirements and classification.
4. EN 13116:2001 — Curtain walling. Resistance to wind load. Performance requirements.
5. EN 1991-1-4:2005+A1:2010 — Eurocode 1. Actions on structures. General actions. Wind actions.
6. EN 1999-1-1:2007 — Eurocode 9. Design of aluminium structures. General structural rules.

Movement figures in this guide are calculated from the stated spans, the standard expansion coefficients and the deflection limits quoted in each table. They are arithmetic, not measurements from a built facade, and the stack joint of every elevation is confirmed against its own structural drawing before fabrication.